The all NTP pyrophosphatase and
The all-α NTP pyrophosphatase and NUDIX families maintain the deoxynucleotide triphosphate (dNTP) pool balance, limit the incorporation of harmful nucleotides, and represent a promising therapeutic approach to develop anticancer and immunosuppressive agents., , , , Among these housekeeping enzymes is the dCTPase enzyme, also known as DCTPP1 or XTP3-TPA (XTP3-transactivated protein A), a pyrophosphatase with high affinity towards modified dNTPs. dCTPase has been linked with cancer stemness, poor clinical prognosis and decreased cytotoxic response to cytidine analogues., , , , In addition to the weak and non-selective ligands and (),, three new Vardenafil HCl Trihydrate of potent inhibitors of dCTPase have been recently reported (, and , )., , In order to find new back-up scaffolds, we screened the ChemBridge DiverSET (50,000 compounds), a successful small-molecule library often used in drug discovery, at a single concentration (10µM) against the purified human dCTPase protein. The screening campaign yielded a 0.4% hit rate at 90% inhibition. After inspection for potential pan assay interference compounds (PAINS), and confirmation of hit identity and purity by LC–MS, we prioritised chemotypes where additional analogues within the library displayed preliminary SAR, and were deemed to be promising starting points for development. After the structural review, we explored the chemotypes described in this letter: triazoles, triazolopyrimidines, triazinoindoles, quinoline hydrazones and arylpiperazines. These chemical scaffolds have previously been investigated by the pharmaceutical industry against unrelated biologic targets, such as mGluR5, PDE2, CB2R, VEGF and 5-HTR (, , , , and , )., , , , In addition, we used DataWarrior software to characterise top hits for their Ligand Efficiency (LE) and Lipophilic ligand Efficiency (LLE), two parameters which estimate the efficiency of a binding interaction in relation to molecular weight and lipophilicity, respectively. Ligand efficiency scores can aid in hit selection and optimization, and LE ≥0.3/LLE ≥3 scores can bias compound selection towards smaller and more specific inhibitors., , The 1,2,3-triazole hit cluster is illustrated by the general formula depicted in , and contained several sub-micromolar inhibitors, with compound as the most active inhibitor in this series (IC=0.36µM). Replacement of the R3 phenyl -chlorine atom in with - or -methyl groups, led to a moderate drop in activity ( IC=0.86µM and IC=1.3µM). Swapping the R1 phenyl group with a butyl chain abolished activity (. ), but a 4-fluoro-phenyl retained sub-micromolar activity ( IC=0.78µM). Replacement of the oxadiazole linker with an amide did not retain potency (. and ), and increasing the bulk at the R2 position, from methyl to cyclopropyl, was an acceptable structural modification (. ). The best compounds in this cluster had a LE of ∼0.35, and LLE of ∼3, deeming them as acceptable starting points for hit optimization (a). The [1,2,4]triazolo[1,5-]pyrimidine cluster followed the general formula in , and revealed compound as the most interesting inhibitor in the series (IC=0.62µM). Replacement of the R1 -methyl substituent in with a -bromo substituent in abolished enzymatic activity. Simultaneous exchange of nitrogen for sulfur at R1, removal of the methyl at R4, and replacement of chlorine for methyl ketone at R3, was detrimental for activity, but still delivered sub-micromolar potency ( IC=0.84µM). An aromatic moiety in R1 was essential for activity, as seen with the inactive ether analogue . Cyclisation of R3 with R4 into a lipophilic ring, such as cyclopentane () or cyclohexanone (), while having R2 as –H or –OH, diminished potency, but still delivered >85% inhibition at 10µM. Replacing R1 with a 3-pyridyl group, with concomitant removal of the methyl at R2 and introduction of an ethyl ester group at R3, delivered sub-micromolar potency ( IC=0.85µM). The 3-pyridyl subset of compounds indicated that larger substituents at R2, i.e. phenyl (), or bulky amide substituents at R3 () cannot be accommodated by the binding site. The best compounds in this cluster had LE ∼0.4 and LLE ∼4, indicating promising binding efficiency (b). In addition, the top [1,2,4]triazolo[1,5-]pyrimidine hits were predicted to have the highest aqueous solubility (cLogS), presenting the most encouraging binding efficiency/solubility profile (f).